Press
Field notes on AI & on-chain data.
Stress-testing an LP position with AI, step by step
A walkthrough of the risk workflow we run for clients: an AI analyst pulls pool reserves and price history, runs 10,000 simulated paths, and explains what a 40% shock would actually do to an ATOM/OSMO position. Then a human checks the math.
Read full storyWhat we learned building an AI governance analyst
We have voted on hundreds of proposals as a validator. This year we built an AI assistant to help us read them. Four iterations later it summarizes proposals with 98% accuracy, but only once we gave it the on-chain parameters and kept a person in the loop.
Read full storyValidator Parity, four years later
In 2022 we showed how skewed Osmosis voting power was. We have re-run the numbers every quarter since. The Nakamoto coefficient has nearly doubled and the top-10 share keeps falling. Here is what changed, and what still has not.
Read full storyGrounding AI in chain data: why tools beat clever prompts
We asked 250 real questions about on-chain data under four setups, from a bare prompt to an AI with live tool access and automated checks. Error rates fell from 38% to 1.5%. The difference was not the prompt; it was where the numbers came from.
Read full storyFive treasury risk metrics every DAO should track
Most DAO treasuries are one bad month away from a runway problem, and most do not know it. These are the five numbers we check first in every treasury review, and how to monitor them automatically.
Read full storyOsmosis in Graphs: The value of an Osmo
Every day, millions of dollars are traded in and out of Osmosis pools. From users compounding their LP rewards to traders exchanging tokens, every transaction slowly yet surely increases the TVL of Osmosis. When the TVL of Osmosis increases, the price of an OSMO tends to follow. In this post, we're going to take a deep dive into the pricing of an OSMO.
Read full storyOsmosis in Graphs: Validator Parity
There is definitely a non-normal spread of voting power among validators. That's not necessarily a bad thing, and this is not a critique of anyone's staking selection or of any validators in general. Currently, there's no incentive to stake with a smaller validator aside from your own desire to decentralize voting power and support individual validators' efforts.
Read full storyImpermanent Loss & You
Osmosis Governance recently approved another set of external incentive matching for the BTSG pools. How do these pools perform if you ignore incentives? Simply swap fees accrued vs impermanent loss. When you put tokens in a pool, you are investing in the GAMM itself.
Read full story